Enhanced biological phosphorus removal fails on many UK works for one reason: not enough volatile fatty acid in the influent to feed the organisms that store phosphorus. Sidestream fermentation makes the shortfall on site, from sludge the works already produces.

The problem sidestream fermentation solves

Phosphorus-accumulating organisms need volatile fatty acids in the anaerobic zone. Take up VFA there, release phosphate; take up phosphate in excess downstream. No VFA, no cycle.

The controlling ratio is VFA-COD to total phosphorus in the influent. Below roughly 10 mg COD per mg P, EBPR is unstable — it works in summer, fails after storm dilution, and the works falls back on ferric dosing with the sludge production and consent risk that brings. Many UK catchments, particularly those with long or well-aerated sewers where fermentation has already happened, sit below that line.

The sidestream answer: ferment a slipstream of primary or waste activated sludge in a dedicated reactor, and feed the VFA-rich liquor to the anaerobic zone. The carbon comes from the works’ own solids rather than from a delivered chemical, and the supply is decoupled from what the sewer happens to deliver that day.

Stopping fermentation at the right stage

Anaerobic digestion runs hydrolysis, then acidogenesis, then methanogenesis. A fermenter wants the middle stage and must actively prevent the third — methanogens would consume exactly the acids being manufactured.

ControlSettingWhy
SRT3–5 daysBelow the roughly 6-day minimum for methanogens at 20 °C, so they wash out
pH5.0–5.5Inhibits methanogens; acidogens tolerate it
Redox−250 to −400 mVAnaerobic, but above the range where sulphate reduction takes over
Temperature>15 °CBelow this the hydrolysis rate makes the fermenter uneconomically large
The SRT window is the primary control and it is narrow. Too short and hydrolysis has not proceeded far enough to release much substrate; too long and methanogens establish and eat the product. Losing pH control has the same effect as losing SRT control.

Yield kinetics

VFA yield approaches a ceiling set by the biodegradable fraction of the feed, at a rate set by hydrolysis:

YVFA = Ymax · [ 1 − exp(−khydt) ]Ymax = 0.20–0.35 mg VFA-COD per mg VS for primary sludge; 0.10–0.20 for waste activated sludge

with the usual temperature correction:

khyd(T) = khyd,20 · θ(T−20)khyd,20 = 0.15–0.25 d−1; θ ≈ 1.08

Worked example — yield at 15 °C. Taking khyd,20 = 0.18 d−1, Ymax = 0.30, SRT = 4 days.
  • Temperature-corrected rate: 0.18 × 1.08−5 = 0.123 d−1.
  • Exponent: 0.123 × 4 = 0.490.
  • Yield: 0.30 × (1 − e−0.490) = 0.30 × 0.387 = 0.116 mg VFA-COD per mg VS fed.

At 20 °C the same fermenter returns about 0.157 — a third more, for nothing but temperature. This is why sidestream fermentation is a summer-strong, winter-weak process and why the fermenter should be sized on the winter case.

Which acids, and why it matters

Not all VFAs are equivalent to the organisms that matter.

AcidPAO uptakeBehaviour
Acetate8–12 mg P per g VSS·hFastest; preferred by Candidatus Accumulibacter
Propionate5–8 mg P per g VSS·hSlower but gives more stable EBPR under variable load; disfavours competing glycogen-accumulating organisms
Butyrate, valerateIndirectConverted to acetate and propionate before uptake

A ratio of roughly two or three parts acetate to one part propionate is a good target: fast enough uptake to work, with enough propionate to hold the population stable when load swings. Primary sludge fermentation typically yields 60–70 per cent acetate and 20–30 per cent propionate, which lands close to that target without intervention. Waste activated sludge yields more propionate, 40–50 per cent, because amino acid fermentation favours it.

The competing organisms and the wider mechanism are covered in biological nutrient removal, and the chemical alternative in chemical phosphorus removal with metal salts.

Sizing the fermenter

Worked example — 100,000 PE works.
  • Primary sludge production at 40 g DS per PE per day: 4,000 kg DS/d.
  • Volatile fraction at 70 per cent: 2,800 kg VS/d.
  • VFA production at Y = 0.28 (20 °C): 2,800 × 0.28 = 780 kg VFA-COD/d.
  • Sludge flow thickened to 4% DS: 4,000 / 40 = 100 m³/d.
  • Fermenter volume at 4-day SRT: 400 m³.

Sense-check against demand: at 100,000 PE and, say, 8 mg/L total phosphorus in roughly 20,000 m³/d, the works receives about 160 kg P/d. At a working requirement near 10 mg VFA-COD per mg P, that is about 1,600 kg COD/d. The fermenter supplies 780 kg — a substantial supplement, not a complete replacement, which is the realistic expectation.

Thickening the feed is what makes the fermenter affordable: at 1 per cent solids the same mass needs 400 m³/d of feed and a 1,600 m³ vessel. See sludge thickening design.

Getting the acids back to the process

The VFA-rich liquor is separated from the fermented solids and returned to the anaerobic zone; the solids go on to digestion. Fermentate typically carries 800–1,200 mg VFA-COD/L. In flow terms it is a small stream — of the order of 100 m³/d against 20,000 m³/d of influent, well under one per cent. Where a larger recycle is quoted, it refers to an elutriation stream used to wash acids out of the solids, not to the fermentate itself.

Design rules

  • Check the influent VFA-COD to TP ratio first. Above about 10 the works probably does not need a fermenter; below it, EBPR will be unstable without one.
  • Hold SRT at 3–5 days and pH at 5.0–5.5. Both exist to keep methanogens out; losing either loses the product.
  • Size on the winter case. Yield falls by about a third between 20 and 15 °C, and winter is when EBPR is already least stable.
  • Thicken the feed to about 4 per cent before fermenting. Vessel volume scales directly with feed flow.
  • Aim for 2:1 to 3:1 acetate to propionate. Primary sludge fermentation lands near this unaided; adding waste activated sludge shifts it towards propionate.
  • Expect a supplement, not a substitute. A realistic fermenter covers roughly half the VFA demand and cuts metal salt dosing by 50–80 per cent rather than eliminating it.
  • Economic above about 50,000 PE where primary settlement and thickening already exist.

Frequently asked questions

How do I know whether my works needs sidestream fermentation?

Measure the influent VFA-COD to total phosphorus ratio across a range of conditions, including after rainfall. Consistently below about 10 mg COD per mg P means EBPR will be carbon-limited and unstable, and a fermenter is the structural fix. Above it, instability is more likely a configuration or return-liquor problem.

Why must the fermenter SRT be so short?

To wash out methanogens, whose minimum SRT is around 6 days at 20 °C. Run the fermenter longer and they establish and consume the acetate you are producing. Three to five days is long enough for useful hydrolysis and short enough to exclude them; low pH provides the second line of defence.

Will a fermenter eliminate ferric dosing?

Usually not. A realistically sized fermenter supplies around half the VFA demand, which typically cuts metal salt dosing by 50–80 per cent and greatly improves stability. Treat it as reducing chemical dependence and sludge production rather than removing them.

Does propionate or acetate give better phosphorus removal?

Acetate gives faster uptake; propionate gives more stable performance under variable load, partly because it disadvantages the glycogen-accumulating organisms that compete with PAOs for carbon. A mixture in the region of two or three parts acetate to one part propionate is the practical target, and primary sludge fermentation produces roughly that unaided.

Sources & further reading